01
Run the flow once
Click through the journey you care about in the recorder, or upload a quick demo video. No selectors, no boilerplate.
MCP documentation & QA, automated
Document and test your MCP server automatically.
Connect your MCP server and Olivia discovers every tool, writes the docs, runs smoke tests with real arguments, and reports bugs with evidence — in minutes, not sprints.
Olivia workspace
Live outputsAPI Docs + QA
GET /users
POST /orders
QA run: 15 tests
MCP Docs + QA
search_docs
create_ticket
Smoke: 8 cases
E2E Recorder + Playwright
NewRun oncecheckout.spec.tsCommitted
2 bugs found with evidence
New — E2E Automation
Run it once. Olivia writes the Playwright test, commits it, and keeps it green.
No more hand-writing brittle end-to-end tests. Walk through a flow a single time and Olivia generates a real Playwright spec with AI, commits it into your repo, and self-heals it as your app evolves.
02
AI writes the Playwright test
Olivia turns what you did into a clean, readable Playwright spec with stable locators and assertions.
03
Commit straight to your repo
The test lands in your project structure with a commit, matching your existing Playwright config and conventions.
04
It heals when the app changes
When selectors drift or flows shift, Olivia repairs the test instead of letting it fail silently.
Try E2E automation Works with your existing Playwright setup.
API Automation
For repositories, endpoints, and API QA.
This side of Olivia is built around code repositories. It scans API routes, generates endpoint documentation, keeps docs aligned with pull request changes, and runs QA against documented endpoints.
- Endpoint documentation
- Methods, paths, params, bodies
- Responses and examples
- Repo backfills
- PR change updates
- Endpoint QA runs
MCP Automation
For servers, tools, schemas, and tool QA.
This side of Olivia is built around running MCP servers. It connects to the server, discovers tools, documents each tool, captures sample behavior, creates smoke suites, and tracks tool-level bugs.
- Tool documentation
- Input schemas and required args
- Sample arguments
- Live tool responses
- Smoke suites per project
- Tool-level bug reports
QA Layer
Testing is not an afterthought. It is generated from what Olivia understands.
API QA and MCP QA do not run the same way. Olivia keeps the workflows separate: endpoint tests for APIs, tool execution tests for MCP, shared reporting for bugs.
Smoke
Fast checks for the paths that must work
For APIs, Olivia tests generated endpoint behavior. For MCP, Olivia generates and runs tool smoke suites with real arguments.
Regression
Reusable coverage for behavior you already trust
Turn documented behavior into repeatable checks so the next release does not quietly break what worked yesterday.
Bugs
Failures written like engineering tickets
Each issue includes the target, scenario, severity, evidence, and recommendation so a developer can fix it without decoding a vague report.
Workflow
Three sources. Three outputs. One review surface.
Olivia is not only a docs generator. It is a workspace for turning engineering systems into readable documentation, runnable checks, and actionable bug reports.
API source
Connect GitHub, select a repository, run a backfill, or let pull request changes update endpoint docs.
MCP source
Connect an MCP server over stdio, HTTP, or SSE and let Olivia inspect available tools and schemas.
E2E source
Record a browser flow once and Olivia generates a Playwright test, commits it to your repo, and heals it over time.
API docs
Generated from repositories, endpoint scans, and pull request changes.
MCP docs
Generated from connected servers, tools, schemas, and sample responses.
Bug evidence
Generated from QA runs with severity, evidence, and recommendations.
Pricing
Simple plans for documentation and QA at any scale.
Take just the half you need — the API side or the MCP side — or both. Every plan includes docs, tests, the bug hunter and the watcher that keeps them current after each merge. Bring your own Claude key and any plan is 40% off with no usage cap.
MCP
$99/month
$59/month with your own Claude key — no usage cap
For teams whose product is an MCP server.
- Up to 3 MCP projects, unlimited tools
- Tool docs, smoke and regression tests
- MCP bug hunter, no monthly cap
- Watcher: re-checks your server after every merge
- ~400 tool docs a month
- The API side is not included
API
$180/month
$108/month with your own Claude key — no usage cap
For teams documenting and testing their own API.
- Up to 3 repositories
- Endpoint docs straight from the code
- Smoke and regression tests per endpoint
- Bug hunter with real requests and evidence
- Watcher: new endpoints documented and tested after every merge
- ~60 watched merges a month
- The MCP side is not included
Pro
$499/month
$299/month with your own Claude key — no usage cap
Both sides, for a team shipping continuously.
- Everything in API and MCP
- Up to 5 repositories, unlimited MCP projects
- UI automation (Playwright) included
- SSO and team roles
- ~200 watched merges a month
- Priority support
Enterprise
$1500/month
$899/month with your own Claude key — no usage cap
For organizations that need scale, control and their own data.
- Everything in Pro
- Unlimited repositories
- ~700 watched merges a month
- Your own database, or your own region
- Dedicated onboarding and SLA
Just want to see it work? Start free with one repository and one MCP project — docs, tests and a few bug hunter runs a month, no card needed.
A “watched merge” is one pull request merged into a branch you watch. Merges that touch many files, or add many endpoints, use more of your monthly allowance than small ones. Run out and Olivia pauses new AI work until the next month — everything already generated stays. Bring your own Claude key and there is no allowance at all.
Built for teams shipping APIs, MCP servers, and integrations
Separate workflows. Shared visibility.
API teams get endpoint docs and API QA. MCP builders get tool docs and MCP QA. Engineering leaders get a single place to see documentation, test coverage, and bugs.